reverse-blocking modular multilevel converter for battery ... · besss employ distributed battery...
TRANSCRIPT
Reverse-blocking modular multilevel converter for batteryenergy storage systems
Xiaofeng YANG1 , Yao XUE1, Bowei CHEN1, Yajie MU1,
Zhiqin LIN1, Trillion Q. ZHENG1, Seiki IGARASHI2
Abstract Energy storage systems with multilevel con-
verters play an important role in modern electric power
systems with large-scale renewable energy integration.
This paper proposes a reverse-blocking modular multilevel
converter for a battery energy storage system (RB-MMC-
BESS). Besides integrating distributed low-voltage batter-
ies to medium or high voltage grids, with the inherited
advantages of traditional MMCs, the RB-MMC-BESS also
provides improved DC fault handling capability. This
paper analyzes such a new converter configuration and its
operating principles. Control algorithms are developed for
AC side power control and the balancing of battery state of
charge. The blocking mechanism to manage a DC pole-to-
pole fault analyzed in depth. Comprehensive simulation
results validate both the feasibility of the RB-MMC-BESS
topology and the effectiveness of the control and fault
handling strategies.
Keywords Reverse blocking, Modular multilevel
converter, Battery energy storage system, SOC control,
Fault blocking
1 Introduction
In recent years, ever-increasing energy demands and
shortage of traditional fossil fuels seriously challenged the
sustainable development of human society. At the same
time, increasing concern about environment problems and
carbon emissions of fossil fuels has provoked worldwide
active research on the next-generation electric power sys-
tem, which is known as the smart grid [1–3] and/or energy
internet (EI) [4, 5]. It features renewable energy resources
and intelligent energy management. The coming energy
power generation system shifts from reliance on fossil fuels
to various renewable energy resources, such as solar and
wind power, etc. [6] However, the stochastic nature of
some renewable resources also brings new challenges to
the reliability and stability of existing power grids. Thus it
is expected that using grid-connected energy storage sys-
tem (ESS) for power buffering, peak shaving, load level-
ling and load frequency control, shall be important for
modern electric power systems with large-scale renewable
energy integration [7, 8]. Available energy storage tech-
nologies include batteries, super capacitors, flywheels, and
pumped hydro storage, where batteries are generally
CrossCheck date: 23 May 2017
Received: 30 June 2016 / Accepted: 23 May 2017 / Published online:
24 June 2017
� The Author(s) 2017. This article is an open access publication
& Xiaofeng YANG
Yao XUE
Bowei CHEN
Yajie MU
Zhiqin LIN
Trillion Q. ZHENG
Seiki IGARASHI
1 School of Electrical Engineering, Beijing Jiaotong
University, Beijing 100044, China
2 Fuji Electric Co., Ltd., Tokyo 141-0032, Japan
123
J. Mod. Power Syst. Clean Energy (2017) 5(4):652–662
DOI 10.1007/s40565-017-0300-5
considered as the dominant new solution for large-scale
ESS due to its ability to supply power for periods of up to a
few hours [9, 10]. The cost of batteries is descending
rapidly and is expected to compete with pumped hydro in
the near future [11].
To integrate battery energy storage system (BESS)
installations to the grid, power converters should have the
following features: (1) fault ride-through capability; (2)
high redundancy and error correction capabilities. For
megawatt-scale medium-voltage (MV) application, multi-
level converters show significant advantages over con-
ventional topologies [12]. Among them, the most
promising concept for renewable energy integration and
power transmission is the modular multilevel converter
(MMC) [13–15]. Compared with conventional multilevel
converters, MMCs provide advantages of high modularity,
better harmonic spectra, lower switching frequency, higher
efficiency, and reduced weight of the filtering
components.
In the last decades, MMCs have attracted the interest of
both academia and industry. The published literature
mainly relates to their applications to high-voltage DC
(HVDC) transmission [16–18], MV electric drives [19–21],
and STATCOMs [22, 23]. MMCs used in BESSs for
interfacing low- or medium-voltage batteries to medium-
or high-voltage grids were reported in [24–27]. They
enable a flexible scaling of power modules to integrate
energy storage and power electronics to a wide range of
operating voltages, output power and stored energy.
However, existing MMC-based BESSs (MMC-BESSs) do
not address DC fault handling capabilities, and special
control issues arising in MMC-BESSs have not yet been
fully overcome. Reverse blocking (RB) IGBTs have a
symmetrical blocking voltage characteristic. Due to can-
celling anti-parallel diodes, the conduction loss of RB-
IGBTs is lower than that of normal IGBTs, so they are
especially suitable for multilevel converters for low
switching frequency application [28].
This paper proposes a reverse blocking MMC-BESS
(RB-MMC-BESS) for enhancing the DC fault handling
capability, which consists of new sub-modules (SMs) and
distributed battery banks. Typical operating principles,
detailed battery energy controls, and fault blocking mech-
anisms are thoroughly analyzed.
The paper is organized as follows. Section 2 describes
the configuration and operating principles of RB-MMC-
BESSs. Section 3 explores the battery energy and state of
charge (SOC) balancing controls in detail. The DC fault
blocking mechanism is analyzed in Section 4. To validate
the feasibility and effectiveness of the proposed topology
and theoretical analysis, extensive simulation results are
discussed in Section 5. Finally, Section 6 reports the main
conclusions.
2 Topologies and basic operations
2.1 Topologies for RB-MMC-based BESSs
A typical (n?1) level three-phase MMC-based BESS is
shown in Fig. 1, comprised of three phase legs where each
leg contains a stack of 2n identical sub-modules (SMs) and
two inductors (La).
For the convenience of discussion, the phase legs are
further divided into an upper arm and a lower arm. Unlike
conventional MMCs, the SMs integrate a battery storage
bank BSM, which can also serve as an active power port.
Fig. 1b illustrates one of the possible SM realizations
proposed in [25], which consists of two IGBTs with
antiparallel diodes and one capacitor that together form a
typical bidirectional chopper. The distributed battery banks
are directly connected across the SM capacitors. However,
when a common DC link short-circuit fault happens, the
diode D2 in each SM will create a fault current path. Large
fault currents cause thermal overstress, which may result in
severe damage to power electronic devices.
The proposed reverse blocking sub-module (RBSM) is
illustrated in Fig. 1c. Unlike the abovementioned SM, two
anti-parallel RB-IGBTs (T2 and T3) are used for the lower
UDC2
a
bc
La
DCP
DCN
SMa1
O
UDC2
SMa2
SMan
SMa(n+1)
SMa(n+2)
SMa(2n)
Usc
Usb
UsaLs
Ls
Ls
isa
isc
isb
La iPc
iNc
iPb
iNb
iPa
iNa
IDC
La
La
La
La
Phase leg Upper/lower arm SMs
SMb1
SMb2
SMbn
SMb(n+2)
SMb(2n)
SMc1
SMc2
SMcn
SMc(n+2)
SMc(2n)
SMb(n+1) SMc(n+1)
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
(b) SM of [25]
X1
X2
CSM
T1
T2
D1
T3
+iSM
USM+Cs
Ts
Rsr
UB
BSM
X1
X2
CSM
T1
T2
D1
D2
UB
BSM
+
(c) Proposed RBSM
(a) Three-phase MMC topology
+
Fig. 1 General configuration of three-phase MMC-BESS
Reverse-blocking modular multilevel converter for battery energy storage systems 653
123
switch, and a bypass circuit consisting of auxiliary thyristor
TS, auxiliary capacitor CS and varistor Rsr is connected in
parallel with the RB-IGBTs. The distributed battery banks
are directly connected across the RBSM capacitors CSM as
before. A MMC-BESS employing RBSMs is hereafter
called a RB-MMC-BESS.
Due to their high degree of modularity, RB-MMC-
BESSs employ distributed battery banks with lower voltage
ratings rather than centralized ones used with a conven-
tional high-voltage common DC link. In case of battery
faults, extra RBSMs can be placed in the phase leg to
replace the damaged ones. RB-MMC-BESSs also have a
fixed common DC link, which may be used to interconnect
them with a MV DC network if desired. What is more, RB-
MMC-BESSs may also transfer power from one phase leg
to another using the controlled internal circulating current
through this common DC link. This is the theoretical basis
for SOC balancing control among phase legs.
2.2 Operating principles of RBSMs
Under normal operations, T2 and T1 operate with com-
plementary switching states, while T3 is switched on all the
time and acts as a free-wheeling diode. Once fault current is
detected, the RBSMs go into fault protection mode by
blocking the control signals of the IGBTs (T1, T2 and T3).
Then the fault current starts to chargeCS throughTS, while Rsr
is designed to prevent over-voltage across CS, thus it further
avoids the potential threats to the main switching devices.
Generally, Rsr is not activated until the voltage of CS
exceeds its threshold value. Thus the function of Rsr will
not be considered in the following to simplify the analysis,
but this does not affect the correctness of the theory. The
switching states of the RBSM are listed in Table 1, where
UB is the battery voltage of BSM and Uclamp is the clamp
voltage across T2 and T3.
2.3 Operation principles of RB-MMC-BESS
As in [29] and without loss of generality, the following
analysis assumes that the operating principles of the three
phases are identical, and the conclusions are taken to apply
to three-phase conditions. Phase-j is taken as an example to
carry out the analysis, where j = a, b, c, and the following
additional assumptions are made:
1) Three-phase AC voltages and currents are pure
sinusoidal and symmetrical.
2) The common DC link voltage UDC is smooth.
3) The AC output current isj is distributed equally
between the upper and the lower legs.
4) Switching losses of the power devices are ignored.
These conditions are not exact for RB-MMC-BESSs but
in general they are fulfilled to a good approximation. The
AC terminals of an RB-MMC-BESS are connected to the
grid Usj through a series connected filter Ls. With reference
direction shown in Fig. 1, the AC currents in normal
operating mode are related by:
iPj ¼1
2isj þ iZj ð1Þ
iNj ¼ � 1
2isj þ iZj ð2Þ
isj ¼ iPj � iNj ð3Þ
iZj ¼1
2iPj þ iNj� �
ð4Þ
where iPj and iNj denote the upper and lower arm currents,
respectively. The arm currents flowing through both the
upper and lower arms consist of half of the AC output
current isj and the common-mode circulating current iZj.
The role of the latter in balancing the SOC of batteries is
discussed below in Section 3.2.
The resulting AC and DC voltages can be calculated as
follows :
uPj ¼UDC
2� uj � La
diPj
dt� RaiPj ð5Þ
uNj ¼UDC
2þ uj � La
diNj
dt� RaiNj ð6Þ
uj ¼uNj � uPj
2� La
2
disj
dt� Ra
2isj ð7Þ
UDC ¼ uPj þ uNj þ 2LadiZj
dtþ 2RaiZj ð8Þ
where uPj and uNj denote the upper and lower arm voltages,
respectively; uj is the AC output phase voltage; UDC is the
rated common DC link voltage; and Ra is the equivalent
series arm resistor.
Under a DC link short-circuit fault condition, if the RB-
MMC-BESS keeps running according to the above rules
before system blocking, UDC will reduce to zero immedi-
ately. Then the inserted RBSMs’ capacitors CSM and bat-
tery banks BSM will be continuously discharged, and
potentially over-discharged if the fault is not cleared
quickly. Therefore some reasonable means for improving
Table 1 Switching states of a RBSM
Operation modes State T1 T2 T3 Ts iSM USM
Normal operation Discharging 1 0 1 1 \0 UB
Charging 0 0 1 1 [0 UB
Bypass 0 1 1 1 – 0
Fault protection Blocking 0 0 0 1 [0 UB
Blocking 0 0 0 1 \0 Uclamp
System starting Charging 0/1 0/1 0/1 0 [0 –
654 Xiaofeng YANG et al.
123
the DC fault ride-through capability of RB-MMC-BESSs
should be found.
3 System controls
A RB-MMC-BESS operates differently to a regular
MMC. Since each RBSM includes its own battery energy
storage, which may act as the DC source, the power is not
only delivered from the common DC link. As previously
described, each arm conducts only half of the AC output
current, thus reducing conduction loss in the converter.
Unbalanced SOC of battery banks may cause premature
failure after extended cycling due to overcharging or
undercharging of batteries. The flat relationship of battery
SOC as a function of their voltage, over a wide range of
voltages, indicates the need for a SOC balancing algorithm
that does not rely on the voltages [30, 31]. Thus SOC
control in RB-MMC-BESSs is one of the main differences
compared to conventional MMCs.
The controller of a RB-MMC-BESS has two main sec-
tions: the power control and SOC balancing control. Fig. 2
shows a block diagram of the system control structure.
3.1 Power control
Active and reactive power control of three-phase RB-
MMC-BESSs is based on decoupled current control.
Considering the sinusoidal output currents, proportional
integration (PI) controllers K1 are adopted in a rotating
frame synchronized with the output frequency. Fig. 3
shows the power control block diagram for a three-phase
RB-MMC-BESS. Here, P* and Q* represent the power
commands for the instantaneous active and reactive power
at the AC side, respectively. The AC side active power P*
causes charging and discharging of the RBSMs’ capacitors
and battery banks, so the SOC and the DC link voltage are
indirectly controlled. Finally, the upper and lower arm
voltage references U�Pj and U�
Nj are determined by the AC
side power.
3.2 SOCs balancing control
The inherent circulating current among phases is required
to charge the capacitors with the lowest SOC and discharge
the ones with the highest SOC. Therefore, it is essential to
control the circulating current of the converter to maximize
the efficiency of the SOC controls. The SOC control struc-
ture of RB-MMC-BESS is illustrated in Fig. 4, where K2 to
K5 refer to close-loop controllers such as PI controllers. SOC
balancing control of the RB-MMC-BESS is divided into
individual SM balancing, phase arm balancing, phase leg
balancing, and inner circulating current control.
Figure 4a shows the block diagram for individual SOC
balancing control; sign() denotes the signum function. This is
responsible keeping all RBSMs in the same arm at the
average arm SOC (i.e. SOCPj, SOCNj) using a close-loop
controller. The average armSOC SOCPj,SOCNj are given by:
SOCPj ¼1
n
Xn
k¼1
SOCjk ð9Þ
SOCNj ¼1
n
X2n
k¼nþ1
SOCjk ð10Þ
The arm SOC balancing control forces the SOC
difference between the upper and lower arms (i.e.
SOCPj-SOCNj) to be zero. The leg SOC control is
zj1
uabcPLL ω P*
Active & reactive power controls
Q*Power control
iabc
SOCjk
Arm SOC balancing
Leg SOC balancing
SOCs balancing control
U*
RB-MMC-BESS main circuitPWM generator
Circulating current control
Averaging
Eq. (14) & Eq. (15)
Individual SOC balancing
Pj
U*Njk,indU*Pjk,indU*j,cir
I*zj2I*zj1
U*Njk
U*Pjk
SOCPj
SOCNj
SOCj,ave
SOCave
SOCPj
SOCNj
U*Nj
Central control system
Enable signal
Fig. 2 Overview of system control structures
Reverse-blocking modular multilevel converter for battery energy storage systems 655
123
designed to force the j-phase average SOC (SOCj;ave) to
follow the average SOC of the three phases (SOCave),
where, SOCj;ave and SOCave are given by:
SOCj;ave ¼1
2n
X2n
k¼1
SOCjk ð11Þ
SOCave ¼1
3
Xc
j¼a
SOCj;ave ð12Þ
These control objectives can be achieved using the
circulating current. Therefore, the output signals of both
leg and arm SOC balancing controllers are i�Zj1 and i�Zj2respectively, from which the reference circulating current
i�Zj is determined:
i�Zj ¼ i�Zj1 þ i�Zj2 ð13Þ
Together, the arm and leg SOC controls result in
direct control of the circulating current in each phase leg,
leading to good current regulation of the battery banks.
The circulating current control loop is illustrated as
Fig. 4. The current minor loop forces izj to follow the
command i�Zj, which generates the voltage control
command U�j;cir.
Finally, the voltage reference for the upper and lower
arm of phase-j is given by:
U�Pjk ¼
U�Pj
nþ U�
Pjk;ind þ U�j;cir þ
UDC
nð14Þ
U�Njk ¼
U�Nj
nþ U�
Njk;ind þ U�j;cir þ
UDC
nð15Þ
where the inputs are shown in Fig. 2.
4 DC fault blocking mechanism
A DC pole-to-pole fault is regarded as one of the most
serious fault types. Therefore, the theory of the DC fault
blocking mechanism will be studied under this condition.
Fig. 5 shows the possible current paths after all IGBTs in a
RBSM are blocked. When current iSM is positive as shown
in Fig. 5a, the capacitor is charged through the anti-parallel
diode D1 and the fault current is limited because the
capacitor voltage UB provides an inverse voltage to switch
off the diode. Otherwise, when current iSM is negative, the
RBSM is bypassed as shown in Fig. 5b. The bypass circuit
goes to work and Cs is charged by the fault current through
the triggered TS. Cs is generally very small compared with
the CSM. Thus ucs will increase quickly to provide the
inverse voltage needed to cut off the arc path at the fault
point.
Once a pole-to-pole DC short-circuit fault occurs, the
common DC link voltage is collapsed to zero and a large
inrush current would be induced. Then the RBSMs will
enter their discharging stage immediately until system
blocking is enabled by the central control system, which is
the same behavior as the equivalent model presented in
[32]. Rather than repeating the detailed explanation found
there, this paper will focus on the fault mechanism in the
blocking stage.
Q*
P*
+ _
+ _
K1
K1
iaibic
idiq
ωLS
ωLS
uaubuc
UdUq
+_
_
++
_
ω
ω
ω
2/3VS
2/3VS
i*q
i*d +
+
+
+
U*dUq*
U*
U*Pj
Nj
dqabc
abcdq
abcdq
Fig. 3 Block diagram of power control
SOCjk
K2+
iPjsign()(k=1,2, ,n)
+SOCPj +
UPjk,ind*
SOCjk
K2+
iNjsign()(k=n+1,n+2, ,2n)
+SOCNj +
UNjk,ind*
(a) Individual SOC control
K3
K4
+ +
+
+
+
+iPj
iNj+
K5
Circulating current controlArm SOC balancing
Leg SOC balancing 0.5
+SOCave
SOCj,ave
izj1*+
izj
+SOCPj
SOCNj
iZj
+Uj,cir*
+
(b) Arm & leg SOC control
*
Fig. 4 Block diagram of SOC control
X1
X2
CSM
T1
T2
D1
T3
+iSM
USM +Cs
Ts
UB
BSM
+
X1
X2
CSM
T1
T2
D1
T3
+iSM
USM +Cs
Ts
UB
BSM
+
(a) iSM (b)0> iSM< 0
Fig. 5 Possible current paths of RBSMs in their blocking state
656 Xiaofeng YANG et al.
123
4.1 Modelling blocking mechanism
Once the RB-MMC-BESS goes into fault protection
mode following a DC pole-to-pole fault, all IGBTs are
blocked by the central control system, and the equivalent of
a phase leg is illustrated in Fig. 6. The equivalent series
capacitance of both the upper and lower arms is expressed
as
Cseq1 ¼ Cseq2 ¼CS
nð16Þ
In this state, the fault current starts to charge the
auxiliary capacitor Cs through Ts. Then ucs increases
quickly to provide the inverse voltage ucsI = 2n9ucs,which
helps to extinguish the fault current. As shown in Fig. 6b, a
second-order oscillating circuit is constructed with the
equivalent series resistance Req, equivalent inductance Leqand the equivalent capacitance Cseq. This will govern the
discharging of Cseq which may be regarded as 2n auxiliary
capacitors Cs in series. The following differential equation
is deduced from Kirchhoff’s voltage law:
d2ucs
dtþ Req
Leq
ducs
dtþ 1
LeqCseq
ucs ¼ 0 ð17Þ
The initial conditions and circuit parameters are:
ucsð0þÞ ¼ ucsð01�Þ ¼ 0
ifð0þÞ ¼ ifð0�Þ ¼ I0
�ð18Þ
Req ¼ 2Ra þ Rf
Leq ¼ 2La
Cseq ¼Cs
2n
8><
>:ð19Þ
where I0 is defined as the initial fault current at the
blocking stage and Rf is the short circuit resistance.
Assuming for simplicity that the auxiliary capacitor
voltages are equal, then the charging current and voltage
of each auxiliary capacitor Cs is:
ucs ¼ e�ts2nI0
xCs
sinðxtÞ ð20Þ
if ¼ �e�tsx0I0
xsin xt � bð Þ ð21Þ
where s is the fault current decay time constant, x0 and xare the natural angular frequency and system angular
frequency, respectively, and b is the initial current phase
angle. These four variables are defined as follows:
s ¼ 4La
2Ra þ Rf
x ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffin
LaCs
� 2Ra þ Rf
4La
� �2s
x0 ¼ffiffiffiffiffiffiffiffiffiffin
LaCs
r
b ¼ arctan
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi16nLa
Cs 2Ra þ Rfð Þ2� 1
s
8>>>>>>>>>>>>><
>>>>>>>>>>>>>:
ð22Þ
Equations (20) and (21) indicate that ucs and the fault
current if are affected by the initial fault current I0; while
ucs is approximately inverse proportional to the RBSM’s
capacitance. In addition, the fault current if is also
influenced directly by the equivalent resistance Req and
inductance Leq.
4.2 Selecting parameters of auxiliary circuit
Without considering system redundancy, it is assumed for
simplicity that the 2n RBSMs are series-connected in each
phase leg. The voltage stress of the auxiliary capacitor always
equals USM under normal operating conditions. After system
blocking is enabled, the auxiliary capacitors are charged in
series by the fault current, and the auxiliary capacitor voltage
will reach its peak value as the current decays to zero. From
(20), the peak blocking voltage across T2 and T3 is
Uclamp ¼ucs;peak
2n¼ e�
bxs
I0
xCs
sinðbÞ ð23Þ
The voltage stress across Cs is illustrated in Fig. 7,
where the clamping voltage is related to both the auxiliary
(a) Charging loop
(b) Simplified equivalent circuit
DCP
isca Phase - b Phase - c
La
La
Cseq1
Ra
+Ts
Cseq2
Ra
+TsDCN
Short circuit
Rf
Fault current charging path
ucs
Req Leq Cseq+
if
Ts
Fig. 6 Current path in the RB-MMC-BESS in the blocking state
Reverse-blocking modular multilevel converter for battery energy storage systems 657
123
capacitance Cs and leg inductance La. A larger leg
inductance value means that more inductive energy will
be transformed to electric field energy, which results in
higher capacitor voltage. The smaller the auxiliary
capacitance and the larger the initial fault current I0, the
faster the capacitor voltage changes.
Therefore the value of Cs is determined by two param-
eters: the voltage limit for IGBTs (UT,max) and the opti-
mized blocking voltage (Ucs,min). The latter helps to
eliminate the AC rectification feeding energy. Therefore,
Cs should satisfy:
Ucs;min � uCs;peak �UT;max ð24Þ
4.3 Fault management
Fig. 8 shows the DC fault protection flow chart for the
RB-MMC-BESS. The operating state is monitored con-
tinuously, with both the DC link voltage and currents sent
back to the central control system, so the DC fault state
may be judged by comparing them with their threshold
values.
Once a DC fault is detected, the system will immedi-
ately block all the trigger pulses in the RB-MMC- BESS to
clear the fault currents. For non-permanent faults, it is
expected that power transmission can be restarted quickly,
so the IGBTs will be triggered to test which type of fault
has occurred. If the fault is cleared then all the IGBTs are
unblocked and the RB-MMC-BESS will be restarted. But if
a permanent fault is identified, both the AC breakers and
the DC breakers are tripped to achieve fault isolation after
fault clearance.
Generally speaking, the fault clearance time achieved
by the protection system is very short, perhaps less than
1ms, to protect the main power devices from thermal
overstress.
5 Verification of RB-MMC-BESS by simulation
To verify the feasibility of the proposed RB-MMC-
BESS and system control strategies, a fully switched sim-
ulation model has been developed as shown in Fig. 9,
where PCC denotes the point of common coupling. The
modulation method adopted in this simulation is the carrier
phase-shifted sinusoidal pulse-width-modulation methods.
Two simulated scenarios are considered in this section. The
first simulation focuses on the system control algorithms
under normal operating conditions, thus demonstrating
control of SOC and power. The second simulation verifies
the DC fault handling capabilities of the RB-MMC-BESS
under a typical DC pole-to-pole fault.
0 0.1 0.2 0.3 0.4 0.5
u Cs,p
eak/U
C
Cs/CSM
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Fig. 7 Voltage stress of CS in the bypass circuit
Start
Normal operation
DC fault
Blocking IGBTs
Y
Y
AC breaker tripped
N
N
Unblocking IGBTs
End
Reclosingpermanent fault?
Fig. 8 Flow chart of DC fault protection in the RB-MMC-BESS
aLa La
La
La
bc
LaLaPCC
UDC
RB-MMC-BESS
Xs/Rs
DCP
DCN
Rf
Pole
-to-P
ole
faul
t
Grid
SM1
SM2
SMn
SM1
SM2
SMn
SM1
SM2
SMn
SM1
SM2
SMn
SM1
SM2
SMn
SM1
SM2
SMn
X1
X2
CSM
T1
T2
D1
T3
iSM
+Cs
Ts
Rsr
UB
BSM
+
Fig. 9 Diagram of the simulated system
658 Xiaofeng YANG et al.
123
5.1 Scenario 1: RB-MMC-BESS under normal
operating conditions
In this scenario, all battery banks in phase a have been
initialized with different SOC. After the RB-MMC-BESS
starts, the power command P* = 2 MW is issued, and all
battery banks in the RBSMs are discharging. The proposed
power control and SOC balancing control algorithms have
been tested and the results are illustrated in Fig. 10.
The convergence SOC curves while discharging is
shown as Fig. 10a. It is clear that the RBSMs with higher
SOC discharge quicker, while some RBSMs with lower
SOC are recharged for a period to bring the SOC closer
together, before they convert to the same value. With the
proposed controller, all the RBSMs’ battery banks are
completely balanced after 9 s. From the power transmission
point of view, the active and reactive power are not
affected by the SOC balancing controls, as shown in
Fig. 10b. Note that the battery parameters are adjusted to
make the simulation time short for this demonstration.
5.2 Scenario 2: RB-MMC-BESS under typical DC
fault condition
Since this paper mainly concerns the DC fault handling
capability of the proposed topology, simulation of a nine-
level RB-MMC-BESS is sufficient to demonstrate the
functionality while maintaining simulation efficiency.
Table 2 summarizes the simulation parameters. The fault
scenario is a non-permanent DC pole-to-pole fault scenario
that occurs at 0.3 s and is cleared at 0.4 s. The simulation
results are shown in Fig. 11 and Fig. 12.
From t = 0 s to 0.3 s, the RB-MMC-BESS operates at a
power rating of 1 MW, supplied by battery banks in the
RBSMs. When the DC pole-to-pole fault occurs at t = 0.3 s,
the common DC link voltage UDC drops to zero immedi-
ately, shown in Fig. 11a. This is accompanied by an inrush
DC short current in the fault point, as seen in Fig. 11b. It is
further supposed that it takes 0.1 ms to block all the trigger
pulses so the RB-MMC-BESS can clear the fault currents.
According to the fault blocking mechanism theory in
Section 4, the 2n series-connected auxiliary capacitor
voltages ucsI provide the inverse voltage that will help to
extinguish the arc fault current in time. Fig. 11c and
Fig. 11d illustrate the grid voltages and currents during the
DC fault. The short circuit fault energy mainly comes from
the DC fault loop shown in Fig. 6. Because the blocking
measures are timely, the grid input currents reduce quickly
to zero, while the grid voltages are not significantly
affected. AC power transmission, shown in Fig. 11e, is
interrupted by the DC fault and recovers very quickly after
the non-permanent fault is cleared.
Fig. 12 illustrates the performance of RBSMs during the
DC pole-to-pole short circuit fault. The SOC and capacitor
voltage of each RBSM is shown in Fig. 12a and Fig. 12b.
Since all IGBTs are blocked in time, the capacitor voltages
remain almost constant at their value at the time of failure,
resulting in the constant SOC of the battery banks. This
maintains a reasonable condition for restarting the RB-
MMC-BESS after the fault is cleared.
Section 3 found that the circulating current can be
controlled according to the desired recharge current of the
battery banks and the equalizing time. Fig. 12c shows the
circulating current during the DC fault. As soon as the
RBSMs go into fault protection mode, the fault current iftransfers to the bypass circuits and starts charging the
auxiliary capacitors Cs through TS. Fig. 12d illustrates the
voltage stress of RBSMs. It can be seen that ucs is directly
applied to the lower switches T2 and T3 when fault pro-
tection mode is enabled.
Table 2 Simulation parameters
System parameters Value
AC line-to-line voltage 7.2 kV
DC link voltage 16.0 kV
RBSM capacitor voltage 2.0 kV
RBSM capacitance 8.2 mF
Leg inductance 3.5 mH
Short circuit resistance (Rf) 5.0 mX
Auxiliary capacitance (Cs) 0.15 lF
-0.5
0.5
1.5
2.5
0 2 4 6 8 10 12
P(M
W)
PQ
0
1.0
2.0
Q(M
var)
-0.5
0.5
1.5
2.5
0
1.0
2.0
t (s)
4 6 8 10 12t (s)
54.4
54.5
54.6
54.7
54.8
54.9
SOC
(%)
0 2
55.0
(a) SOCs of the battery banks
(b) Power transferred
Fig. 10 Convergence of SOC of the simulated battery banks
Reverse-blocking modular multilevel converter for battery energy storage systems 659
123
Fig. 13 compares the simulated DC fault handling
capabilities of a RB-MMC-BESS and a traditional MMC-
BESS system. Once DC short circuit fault occurs in the
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
-5
0
5
10
15
20U
DC
(kV
)
-0.2
0
0.2
0.4
0.6
Faul
tcur
rent
(kA
)
-0.20
0.20.40.6
0.295 0.300 0.305 0.310
Grid
volta
ge(k
V)
-8-6-4-202468 usa usbusc
-0.3-0.2-0.1
00.10.20.3
Grid
curr
ent(
kA)
isaisb isc
-2.5-2.0-1.5-1.0-0.5
00.51.01.5
PQ
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
-2.5-2.0-1.5-1.0-0.500.51.01.5
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
P(M
W)
Q(M
var)
(a) Common DC link voltage
(b) DC fault current
(c) Grid voltage
(d) Grid current
(e) AC power transmission
Fig. 11 Simulation results for RB-MMC-BESS during a DC fault
SOC
(%)
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
54.999454.999554.999654.999754.999854.9999
55.000155.000255.000355.0004
1.996
1.998
2.000
2.002
2.004
USM
,a(k
V)
0.29 0.30 0.31-0.20
-0.15
-0.10
-0.05
0
0.05
i Z abc
(kA
)
iZaiZb
0.299 0.300 0.301 0.302-0.2
-0.1
0
0.1
-1.0-0.5
00.51.01.52.02.5
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
Vol
tage
ofT 2
,3(k
V)
55.0000
t (s)0.20 0.25 0.30 0.35 0.40 0.45 0.50
0.32
iZc
t (s)
(a) SOCs
(b) Capacitor voltages
(c) Circulating currents
(d) Voltage stress of T2 andT3
RBSM capacitor voltage of upper arm
RBSM capacitor voltage of lower arm
Fig. 12 Simulation results for RBSMs during a DC fault
660 Xiaofeng YANG et al.
123
RB-MMC-BESS and all IGBTs are blocked, the series-
connected Cs voltage increases quickly to provide an
inverse voltage to suppress the fault current. However, due
to the lack of fault current suppression measures, the fault
current in the traditional MMC-BESS increases dramati-
cally. The peak value may reach nearly 18 times that of the
RB-MMC-BESS under the same conditions. It is evident
that the RB-MMC-BESS has better fault blocking capa-
bilities, and this one of its salient merits.
6 Conclusion
This paper has investigated the operation and control of
a proposed reverse-blocking modular multilevel converter
with a distributed battery energy storage system (RB-
MMC-BESS) for interfacing low-voltage batteries to the
medium or high voltage grids. Its theoretical performance
has been analyzed and the findings have been confirmed
through simulation.
Unlike conventional MMC-BESS designs, sub-modules
with integrated battery banks use two anti-parallel RB-
IGBTs and an additional bypass circuit. The proposed
design can block fault currents effectively with reduced
requirement for precise trigger pulses during fault condi-
tions. This greatly enhances the ability of the BESS to
respond to fault conditions and to ride through non-per-
manent faults.
The RB-MMC-BESS also employs direct management
of state of charge (SOC) of battery banks, rather than sub-
module voltages. Control algorithms combining power
control and SOC control have been developed and
demonstrated through simulation. The standard modulation
strategies can be employed under the normal operating
condition.
Acknowledgements This work was supported by the State Key
Laboratory of Large Electric Drive System and Equipment Tech-
nology (No. SKLLDJ042016005), the National Key Research and
Development Program of China (No. 2016YFE0131700), and the
National Natural Science Foundation of China (No. 51577010). The
authors would also like to thank you FUJI Electric Co., Ltd. for
providing the technical datasheet of RB-IGBTs modules.
Open Access This article is distributed under the terms of the
Creative Commons Attribution 4.0 International License (http://
creativecommons.org/licenses/by/4.0/), which permits unrestricted
use, distribution, and reproduction in any medium, provided you give
appropriate credit to the original author(s) and the source, provide a
link to the Creative Commons license, and indicate if changes were
made.
References
[1] Gungor VC, Sahin D, Kocak T et al (2013) A survey on smart
grid potential applications and communication requirements.
IEEE Trans Ind Inform 9(1):28–42
[2] Wang J, Huang AQ, Sung W et al (2009) Smart grid tech-
nologies. IEEE Ind Electron Mag 3(2):16–23
[3] Fang X, Misra S, Xue G et al (2012) Smart grid—the new and
improved power grid: a survey. IEEE Commun Surv Tutorials
14(4):944–980
[4] Huang AQ, Crow ML, Heydt GT et al (2011) The future
renewable electric energy delivery and management (FREEDM)
system: the energy internet. Proc IEEE 99(1):133–148
[5] Sun Q, Zhang Y, He H et al (2017) A novel energy function-
based stability evaluation and nonlinear control approach for
energy internet. IEEE Trans Smart Grid 8(3):1195–1210
[6] Whittingham MS (2012) History, evolution, and future status of
energy storage. Proc IEEE 100(Special Centennial
Issue):1518–1534
[7] Dell RM, Rand DAJ (2001) Energy storage—a key technology
for global energy sustainability. J Power Sources 100(1–2):2–17
[8] Koller M, Borsche T, Ulbig A et al (2015) Review of grid
applications with the Zurich 1 MW battery energy storage sys-
tem. Electr Power Syst Res 120:128–135
[9] Divya KC, Østergaard J (2009) Battery energy storage tech-
nology for power systems—an overview. Electr Power Syst Res
79:511–520
[10] Zidar M, Georgilakis PS, Hatziargyriou ND et al (2016) Review
of energy storage allocation in power distribution networks:
applications, methods and future research. IET Gener Transm
Distrib 10(3):645–652
[11] Lawder MT, Suthar B, Northrop PWC et al (2014) Battery
energy storage system (BESS) and battery management system
(BMS) for grid-scale applications. Proc IEEE 102(6):1014–1030
[12] Abu-Rub H, Holtz J, Rodriguez J et al (2010) Medium-voltage
multilevel converters - state of the art, challenges, and require-
ments in industrial applications. IEEE Trans Ind Electron
57(8):2581–2596
[13] Marquardt R, Lesnicar A (2004) New concept for high volt-
age—modular multilevel converter. In: PESC 2004 conference,
Aachen, Germany, p 5
[14] Dorn J, Huang H, Retzmann D (2007) Novel voltage-sourced
converters for HVDC and FACTS applications. In: Proceedings
of CIGRE symposium, Osaka, Japan, p 8
[15] Adam GP, Ahmed KH, Finney SJ et al (2013) New breed of
network fault-tolerant voltage-source-converter HVDC trans-
mission system. IEEE Trans Power Syst 28(1):335–346
HBSM-MMC-BESSRB-MMC-BESS
023.00.3000.295t (s)
-2
0
2
4
6
Faul
tcur
rent
(kA
)
8
10
0.305 0.310 0.315
AC feeding effect
Fig. 13 Fault currents of a RB-MMC-BESS and a traditional MMC-
BESS
Reverse-blocking modular multilevel converter for battery energy storage systems 661
123
[16] Saeedifard M, Iravani R (2010) Dynamic performance of a
modular multilevel back-to-back HVDC system. IEEE Trans
Power Deliv 25(4):2903–2912
[17] Moon JW, Park JW, Kang DW et al (2015) A control method of
HVDC-modular multilevel converter based on arm current
under the unbalanced voltage condition. IEEE Trans Power
Deliv 30(2):529–536
[18] Trinh NT, Zeller M, Wuerflinger K et al (2016) Generic model
of MMC-VSC-HVDC for interaction study with AC power
system. IEEE Trans Power Syst 31(1):27–34
[19] Antonopoulos A, Angquist L, Norrga S et al (2014) Modular
multilevel converter AC motor drives with constant torque from
zero to nominal speed. IEEE Trans Ind Appl 50(3):739–746
[20] Hagiwara M, Hasegawa I, Akagi H (2013) Startup and low-
speed operation of an electric motor driven by a modular mul-
tilevel cascade inverter (MMCI). IEEE Trans Ind Appl
49(4):1556–1565
[21] Jung JJ, Lee HJ, Sul SK (2015) Control strategy for improved
dynamic performance of variable-speed drives with modular
multilevel converter. IEEE J Emerg Sel Top Power Electron
3(2):371–380
[22] Du S, Liu J (2013) A study on DC voltage control for chopper-
cell-based modular multilevel converters in D-STATCOM
application. IEEE Trans Power Deliv 28(4):2030–2038
[23] Hm P, Bina MT (2011) A transformerless medium-voltage
STATCOM topology based on extended modular multilevel
converters. IEEE Trans Power Electron 26(5):1534–1545
[24] Hagiwara M, Akagi H (2014) Experiment and simulation of a
modular push–pull PWM converter for a battery energy storage
system. IEEE Trans Ind Appl 50(2):1131–1140
[25] Quraan M, Yeo T, Tricoli P (2016) Design and control of
modular multilevel converters for battery electric vehicles. IEEE
Trans Power Electron 31(1):507–517
[26] Vasiladiotis M, Rufer A (2015) Analysis and control of modular
multilevel converters with integrated battery energy storage.
IEEE Trans Power Electron 30(1):163–175
[27] Soong T, Lehn PW (2014) Evaluation of emerging modular
multilevel converters for BESS applications. IEEE Trans Power
Deliv 29(5):2086–2094
[28] Igarashi S, Miyashita S, Wakimoto H et al (2011) Advanced
three-level converter with newly developed 1200 V reverse
blocking IGBTs. In: Proceedings of 2011–14th european con-
ference on power electronics and applications (EPE 2011),
Birmingham, UK, 30 Aug–1 Sept 2011, p 7
[29] Yang X, Li J, Wang X et al (2011) Circulating current model of
modular multilevel converter. In: Proceedings of 2011 Asia-
Pacific power and energy engineering conference (APPEEC),
Wuhan, China, 25–28 March 2011, p 6
[30] Maharjan L, Inoue S, Akagi H et al (2009) State-of-charge
(SOC)-balancing control of a battery energy storage system
based on a cascade PWM converter. IEEE Trans Power Electron
24(6):1628–1636
[31] Li B, Shi S, Wang B et al (2016) Fault diagnosis and tolerant
control of single IGBT open-circuit failure in modular multilevel
converters. IEEE Trans Power Electron 31(4):3165–3176
[32] Han P, Wang S (2012) Parameter coordination of modular
multilevel converter for robust design during DC pole to pole
fault. In: Proceedings of 2012 China international conference on
electricity distribution (CICED), Shanghai, China, 10–14 Sept
2012, p 4
Xiaofeng YANG received the B.S. and Ph.D. degrees from Beijing
Jiaotong University, Beijing, China, in 2003 and 2011, respectively,
all in electrical engineering. Currently, he is a Lecture of Beijing
Jiaotong University. From 2003 to 2005, he was a power supplier
engineer of the seventh Research Institute of China Ministry of
Information Industry, Guangzhou, Guangdong. From 2012 to 2014,
he was a Postdoctoral Fellow at North Carolina State University,
Raleigh, NC. His current research interests include HVDC, high
power energy converter, power quality control.
Yao XUE received his B.S. degree in electrical engineering from
Beijing Jiaotong University, Beijing, China, in 2014; and is currently
working toward the Ph.D. degrees in electrical engineering from
Beijing Jiaotong University. His current research interests are
modular multilevel converter and VSC-HVDC.
Bowei CHEN received his B.S. degree in electrical engineering from
Qingdao University of Science and Technology, Shandong, China, in
2015, and is presently a graduate student at Beijing Jiaotong
University. His current research interest is the modular multilevel
converter.
Yajie MU received her B.S. degree in electrical engineering from
Beijing Jiaotong University, Beijing, China, in 2015, and is presently
a graduate student at Beijing Jiaotong University. Her current
research interest is the reverse blocking IGBTs based converter.
Zhiqin LIN received the B.S. degree in electrical engineering and
automation from Beijing Jiaotong University, Beijing, China in 2011,
where he is currently working toward the Ph.D. degree in the School
of Electrical Engineering from 2013. His research interests include
multilevel converter, power quality management.
Trillion Q. ZHENG received the B.S. degree in electrical engineer-
ing from Southwest Jiaotong University, Sichuan, China, in 1986 and
the M.S. and Ph.D. degrees in electrical engineering from Beijing
Jiaotong University, Beijing, China, in 1992 and 2002, respectively.
He is currently the Distinguished Professor at Beijing Jiaotong
University.
Seiki IGARASHI received the B.E. and M.E. degrees in electrical
engineering from Meiji University, Japan, in 1982 and 1984,
respectively. In 1984, he started working at Fuji Electric Corporate
R&D Center. He was responsible for development of the high
efficiency Fuel Cell Inverter, UPS and Industrial Power supplies.
From 2003, he moved to the Semiconductor Group. Now he is Senior
Manager for Device Application Technology Department Fuji
Electric Co., Ltd., Japan. He is member of IEE Japan. He received
an Excellent Paper Award from IEE Japan in 2000. His research
interests include new power device development planning and its
application technologies.
662 Xiaofeng YANG et al.
123